Xref: utzoo comp.ai:9544 comp.ai.philosophy:1059 Newsgroups: comp.ai,comp.ai.philosophy Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!van-bc!ubc-cs!yking From: yking@cs.ubc.ca (Yossarian Yggy King) Subject: Visual stability and perception/cognition Message-ID: <1991Jun25.223559.26962@cs.ubc.ca> Sender: yking@cs.ubc.ca (Yossarian Yggy King) Organization: University of British Columbia, Vancouver, B.C., Canada Date: Tue, 25 Jun 91 22:35:59 GMT 25 June 1991 Stability of the visual world, allowing us a coherent perception of the world from a continually fluctuating retinal image, is the result of our visual system integrating information over some period of time. At a recent spatial vision conference at York University, the issue of visual stability came up during one of the question periods. I forget the exact course the discussion followed, but it centered mainly on stability of the visual world across very small periods of time, on the order of the interval between eye movements. It was argued by, among others, Dana Ballard, that a shift from retinal coordinates to object-based coordinates provides a natural mechanism for imposing stability. While small scale integration over time is certainly an issue that needs to be addressed, I would say that similar issues arise at many temporal scales. We have a clear conscious impression of our environment that remains consistent and stable across periods of time greatly extended beyond periods of eye fixation. Over longer scales, consideration of "scene based coordinates" seems more reasonable--we are aware of our position, and the position of various objects, in relation the scene as a whole, rather than imposing new coordinate frames for every seperate object. The notions of stability and integration over time, in which past experiences affect present behaviour of a system, can be considered a form of memory. Short term stability of the visual world may only involve a sort of sensory memory, arising from summation over time of populations of neurons in the visual system. In contrast, stability of our "awareness" of "the world out there" brings elements of long term memory into play that allow us to remember what room we are in, what objects are behind us, what things look like while our eyes are closed, etc etc. My point in raising these issues is that there appears to be a continuum of world stability--placement of a sharp dividing line between "image stability" and "world stability" seems an impossible task. To push this point to an extreme, the idea of stability can be invoked in the context of spatial "maps." We have stable short term maps of our immediate environment, longer-lasting maps of the area in which we are operating (say, perhaps, a shopping mall) and then seemingly eternal spatial memories of places like the home we grew up in. These extreme cases can perhaps be referred to as visual memory only fancifully, but I think they emphasise that there is at best only a blurry distinction between visual stability and more "traditional" views of memory. I feel that this whole issue is just one example of how unclear the boundary is between perception and cognition. At what point do we stop "seeing" something and start "thinking" about it? Comments, anyone? -- straeH fo neeuQ eht - "!!daeh reh htiw ffO"